JP2001338304A - Device and method for pattern inspection, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform a comparing inspection on an inspecting object pattern image and a reference pattern in real time. SOLUTION: The first edge is detected from the image of an inspecting object pattern. By comparing the first edge with the edge of the first reference pattern, the matching of an inspecting object pattern image with the reference pattern are performed. The first reference pattern is shifted by using an obtained shift quantity S1. By comparing the first edge with the edge of the shifted first reference pattern, an inspecting object pattern is inspected. As one of pattern deforming quantities, a shift quantity S2 is obtained. The second reference pattern is shifted by the portion of shift quantities S1+S2. By using the shifted reference pattern, a profile is obtained on the inspecting object pattern image to detect the second edge. By comparing the second edge and the edge of the shifted second reference pattern, the inspecting object pattern is inspected. A shifting quantity S3 is obtained as one of the pattern deforming quantity.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern inspection apparatus, a pattern inspection method, and a recording medium, and more specifically, for example, to a semiconductor (LSI) or a liquid crystal panel and a mask thereof (LSI) prepared according to design data. The present invention relates to a pattern inspection apparatus, a pattern inspection method, and a recording medium for inspecting a fine pattern such as a reticle.

[0002]

2. Description of the Related Art A pattern inspection of a wafer in a manufacturing process of a semiconductor integrated circuit or a pattern inspection of a mask for forming the pattern is performed by die to die.
e) An optical pattern inspection device using a method called comparison is used. This inspection method is a method of finding a defect by comparing images obtained from the same position of a die to be inspected and a nearby die.

[0003] On the other hand, for inspection of a mask called a reticle having no adjacent die, a die-to-database (d
ie to database) A method called comparison is adopted. That is, a method is used in which CAD data is converted into an image format and used as a substitute for the proximity die to perform the same inspection as described above. The technology is disclosed, for example, in US Pat.
No. 3702 “Automated photomask inspection apparat
However, in this method, the rounded portion of the corner of the actual pattern formed on the wafer is recognized as a defect. As a countermeasure, a rounded image is obtained from the CAD data. In such a situation, a die-to-database comparison inspection would identify pattern deformations that do not need to be determined as corner failures, as defects. However, even if the pre-processing described above is performed, a considerable dilemma arises in that, if the pattern deformation at the corner is set to be ignored, a minute defect existing outside the corner cannot be recognized.

[0004] At present, inspection by a die-to-database comparison method has been put to practical use because a mask must exactly match CAD data. However, the pattern transferred to the wafer is allowed to be deformed within a range in which electrical characteristics and the like are guaranteed, and the pattern is actually deformed to a considerable extent due to a difference in exposure conditions and the like.

In the above-described pattern inspection method of the die-to-die comparison method, a defect called a systematic defect, which is commonly generated in all dies on a wafer due to a mask defect or the like, cannot be detected. That is,
This is because the same defect has occurred in both the die to be inspected and the adjacent die to be compared, and a difference cannot be recognized by comparing the two.

Therefore, although there is a problem in terms of calculation cost and the like, and it has not been put to practical use, a matching inspection between CAD data and a wafer image has been proposed. For example, NEC
Technical Report Vol. 50 No. 6/1997, “Automatic Failure Location Tracing Method for Logic LSI Using Electron Beam Tester”. This document describes a method using projection of wiring edges on the X and Y axes, a method focusing on wiring corners, and a method applying a genetic algorithm. Further, as a method adopted in this document, a matching method is described in which a closed region is extracted after linearly approximating an edge, and the closed region is used. However, none of these methods can achieve a speed that can be used for high-speed inspection, and furthermore, it is not possible to perform matching while detecting the amount of pattern deformation.

At present, an automatic defect type classification (Auto Defect Clas) is performed by comparing an image including a defect (defect image) and an image of a proximity die corresponding thereto (a reference image).
sification: ADC). However,
The uneven brightness of the reference image affects the recognition accuracy. In some cases, it is impossible to specify the inside and outside of the pattern only from the image. In such a case, it is often difficult to distinguish between a short circuit and a defect. Further, since it is not possible to obtain information on which pattern the defect has destroyed, it is not possible to classify a fatal defect to the pattern and a defect that is not.

In the inspection method using the die-to-die comparison, the position of a defect has an error due to the stage accuracy and the optical system accuracy of the inspection device, and the error is about 10 times or more larger than the wiring pattern width. large. Due to this, even if a defect position is projected on a pattern (design pattern) to be formed, the defect position of the pattern cannot be specified accurately.

[0009]

In recent years, the pattern width of an integrated circuit has become about the same as or shorter than the wavelength of a light source used in an exposure process. (Optical Proximity
A method of adding a Correction (OPC) pattern is employed. This is a technique in which a mask is formed by adding an OPC pattern to design data, and exposure is performed using the mask, so that an actual pattern on a manufactured wafer approaches the design data.

Whether the OPC pattern is effectively acting as a correction on the pattern on the wafer is determined by the conventional die
Inspection is not possible with a two-die comparison. Therefore, there is a need for a solution to this problem, for example, a method capable of comparing and verifying the pattern of a wafer with design data in consideration of the allowable pattern deformation amount.

Further, for example, a system-on-chip (SO)
In the high-mix low-volume production seen in C), a short delivery time is required. In such a case, even if a systematic defect is found by an electrical inspection as a final inspection, a short delivery time may not be met. As a countermeasure, there is a demand for monitoring a difference from design data at each stage of the exposure process. Therefore, there is a demand for an inspection method in which pattern deformation that does not affect the electrical characteristics is set as an allowable pattern deformation amount, and design data and a wafer pattern can be compared and verified while considering deformation within the allowable pattern deformation amount.

At present, as a pattern deformation evaluation, a design check is performed by a litho simulator or the like. In order to verify the validity of the simulation, a means for comparing and examining the pattern (simulation pattern) output by the litho simulator and the actual pattern is required.

Further, it has become more and more important to improve the circuit design technology by obtaining the pattern deformation amount with respect to the design data.

At present, CD-S is used for managing the pattern line width of a wafer in a semiconductor integrated circuit manufacturing process.
EM (Critical Dimension Scanning Electron Microsc
ope) is used. The CD-SEM automatically measures the line width of a linear pattern at a designated position using a line profile for each transfer unit of a stepper called a shot. This length measurement was performed at several locations for several shots on several wafers per lot, and whether the transfer function of the stepper was normal, n
It can be managed in m units.

In managing the circuit pattern, in addition to the line width, the shrinkage of the wiring end, the position of the isolated pattern, etc. are also important. Cannot measure. Therefore, the measurement of these two-dimensional shapes is performed by the operator visually observing images obtained from a CD-SEM or another microscope.

Optical proximity correction (OPC) plays an important role not only in securing the line width of a linear pattern but also in forming the shape of a corner or an isolated pattern. Furthermore, due to the improvement of the operating frequency, it is now important to control the shape of the tip and the root of the gate wiring pattern called an end cap or a field extension in addition to the gate line width.

The shape measurement of such a two-dimensional pattern is performed by
This is important both in the sampling inspection in the manufacturing process and in the prototype stage. In particular, in the trial stage, an inspection of the pattern formation is required for the entire surface of the wafer.

However, as described above, management of the two-dimensional shape is performed by human work, and automation is required from the viewpoint of accuracy and productivity.

An object of the present invention is to perform a comparison inspection between a pattern image to be inspected and a reference pattern in real time.

Another object of the present invention is to perform matching while allowing a shape difference within an electrically allowable range.

Another object of the present invention is to perform stable defect detection.

Still another object of the present invention is to enable quantitative and high-speed automatic measurement of a two-dimensional pattern (pattern image to be inspected) which has been performed visually so far.

[0023]

In order to achieve the above object, according to a first aspect of the present invention, there is provided a pattern inspection apparatus for inspecting a pattern to be inspected by comparing the pattern to be inspected with a reference pattern. Storage means for storing the reference pattern, input means for inputting an image of the inspection target pattern, and comparing an edge of the input image of the inspection target pattern with an edge of the stored reference pattern. The inspection means includes an inspection means for inspecting the pattern to be inspected, and an output means for outputting a result of the inspection.

Here, the inspection means may perform matching between the inspection target pattern image and the reference pattern by comparing an edge of the inspection target pattern image with an edge of the reference pattern. it can.

Here, the matching may be performed by expanding an edge of the pattern image to be inspected.

Here, the matching may be performed by expanding an edge of the reference pattern.

Here, the matching may be performed by using, as an evaluation value, a sum of products of the amplitude of the edge of the pattern image to be inspected and the amplitude of the edge of the reference pattern at each pixel.

Here, the matching is performed by using, as an evaluation value, a sum of inner products of edge vectors of the pattern image to be inspected and an edge vector of the reference pattern at each pixel or a sum of absolute values of the inner products. The vector may have the amplitude of the edge as its magnitude and the direction of the edge as its direction.

Here, the matching may be performed by changing a weight for each part of the reference pattern.

Here, the inspection means may associate the edge of each pixel of the reference pattern with the edge of each pixel of the actual pattern image.

Here, the association may be performed in consideration of a distance between an edge of each pixel of the reference pattern and an edge of each pixel of the pattern image to-be-inspected, and a direction difference between both edges. it can.

Here, the inspection means may form an area based on the edge of the pattern image to be inspected which could not be associated, and recognize the area as a defective area.

Here, the inspection means forms an area based on the edge of the pattern image to be inspected which has been able to be associated, and detects an area in which the distribution of luminance is not uniform among the areas. However, the area can be recognized as a defective area.

Here, the inspection means may determine a defect type based on a geometric feature of the defect area.

Here, the inspection means may determine a defect type based on a characteristic amount relating to luminance of the defect area.

Here, the inspection means may calculate a pattern deformation amount of the inspection target pattern with respect to the reference pattern.

Here, the pattern deformation amount includes at least one of a position shift amount, a magnification change amount, and a line width increase amount.
One may be included.

Here, the inspection means may add a pattern attribute to the reference pattern.

Here, the inspection means takes a profile on the pattern image to be inspected, detects a predetermined point for each profile, performs curve approximation on the detected point, and performs edge approximation on the image of the pattern image to be inspected. It can be said that.

According to a second aspect of the present invention, a pattern inspection method according to the present invention is a pattern inspection method for inspecting a pattern to be inspected by comparing the pattern with a reference pattern. And an inspection step of inspecting the inspection target pattern by comparing an edge of the input image of the inspection target pattern with an edge of the reference pattern stored in storage means, and outputting a result of the inspection. And an output step.

According to a third aspect of the present invention, a recording medium according to the present invention is a computer-readable recording medium storing a program for causing a computer to execute a pattern inspection method for inspecting a pattern to be inspected by comparing the pattern with a reference pattern. In the pattern inspection method, an input step of inputting an image of the inspection target pattern, and comparing an edge of the input image of the inspection target pattern with an edge of the reference pattern stored in a storage unit Accordingly, there is provided an inspection step of inspecting the inspection target pattern, and an output step of outputting a result of the inspection.

According to the above configuration, the comparison inspection between the pattern image to be inspected and the reference pattern can be performed in real time.

Further, matching can be performed while allowing a shape difference within an electrically allowable range.

Further, stable defect detection can be performed.

Further, the measurement of the two-dimensional pattern (pattern image to be inspected), which has been performed visually, can be quantitatively and automatically performed at high speed.

[0046]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a diagram showing an example of a theoretical pattern based on design data, and FIG. 2 is a diagram showing an example of a pattern (actual pattern) actually manufactured based on the design data. As shown in FIG. 2, the actual pattern has a short-circuit defect, a defect due to particle adhesion, or a deformation within an allowable deformation amount. Therefore, it differs somewhat from the pattern theoretically obtained based on the design data.

The pattern inspection apparatus according to the present embodiment uses an inspection target pattern (for example, a pattern as shown in FIG. 2).
Is compared with a reference pattern (for example, a pattern as shown in FIG. 1).

FIG. 3 is a diagram showing an outline of the inspection processing performed by the pattern inspection apparatus according to the present embodiment. In the inspection processing, first, a first edge is detected from the image of the inspection target pattern. Next, by comparing the first edge with the edge of the first reference pattern, matching between the inspection target pattern image and the reference pattern is performed. As a result of the matching, the shift amount S ₁ is obtained, the shift amount S ₁
Is used to shift the first reference pattern. Then, the inspection target pattern (actual pattern) is inspected by comparing the first edge with the shifted edge of the first reference pattern. In the first inspection, the amount of pattern deformation is obtained, and a defect is detected. Shift amount S ₂ is obtained as one of the pattern deformation quantity.

Next, in order to detect the second edge from the image of the pattern to-be-inspected, the corresponding second reference pattern is shifted by the shift amount S ₁ + S ₂ . Using the shifted second reference pattern, a profile is obtained on the inspection target pattern image, and the second edge is detected. Then, the inspection target pattern is inspected by comparing the second edge with the shifted edge of the second reference pattern. Also in the second inspection, the amount of pattern deformation is obtained or a defect is detected. Shift amount S ₃ is obtained as one of the pattern deformation quantity.

FIG. 4 is a diagram showing an example of a hardware configuration of the pattern inspection apparatus according to the present embodiment. The pattern inspection apparatus according to the present embodiment includes a main control unit 1, a storage device 2,
Input / output control unit 3, input device 4, display device 5, printing device 6
And an image generation device 7.

The main control unit 1 is constituted by a CPU or the like, and controls the whole apparatus as a whole. The main controller 1 includes a storage device 2
Is connected. The storage device 2 can take the form of a hard disk, a flexible disk, an optical disk, or the like. The main control unit 1 also has an input / output unit 3 such as a keyboard and a mouse, input data,
A display device 5 such as a display for displaying calculation results and the like and a printing device 6 such as a printer for printing calculation results and the like are connected.

The main control unit 1 includes a control program such as an OS (Operating System), a program for pattern inspection,
And an internal memory (internal storage device) for storing required data and the like, and pattern inspection is realized by these programs and the like. These programs can be stored in a floppy (registered trademark) disk, a CD-ROM, or the like, and can be read and executed in a memory, a hard disk, or the like before execution.

FIG. 5 is a functional block diagram of the pattern inspection apparatus according to the present embodiment. Reference pattern generation unit 11, search unit 12, output unit 13, and defect recognition unit 1
4 is realized by a program. The backbone database 21, the recipe database 22, and the defect type reference database 23 are provided in the storage device 2.

The backbone database 21 is provided outside, and the pattern inspection apparatus is connected to the backbone database 21 via the LAN.
May be accessed.

(Recipe) Before the inspection, first, a set of inspection parameters called a recipe is set. The parameters include the distance (pixel interval) between pixels on the actual pattern at the time of image acquisition of the image of the inspection target pattern to be inspected, and 512 × 512 or 1
There are pixel numbers such as 024 × 1024. From these values, the distance (image size) of the image to be processed at a time on the actual pattern can be grasped. In addition, parameters for edge detection and parameters for recognizing defects are set.

As data to be compared with the pattern image to be inspected, design data is used. As the design data, for example, CAD layout data in GDS format is used.
You can use Leia fusion and fracturing. In the present embodiment, a bundle of line segments obtained by this processing is
A rectangular area having a length of one side obtained by adding an error amount of the stage and a maximum parallel movement amount of the pattern to the image size is used as a reference pattern, and the recipe database 2
2 in advance. If the stage error is negligible compared to the maximum parallel movement amount of the pattern, the absolute coordinate value of the pattern deformation can be measured. In the present embodiment, in consideration of the error of the stage and the maximum amount of parallel movement of the pattern, the reference pattern is processed to be larger than the pattern image to be inspected. May be processed.

A process for rounding a corner may be performed on the reference pattern. As shown in FIG.
The design data is a polygon having an acute angle (dotted line in the figure), while the actually formed circuit pattern has rounded corners. Therefore, a circle, an ellipse, a straight line, or a curve described by another method may be applied to the corner portion, and the correction may be made so as to be close to the actual pattern.

If the design data is used as the reference pattern, it becomes a defect inspection for performing a comparison inspection with a pattern to be realized. In this case, an allowable amount that does not affect the electrical characteristics is set as the allowable pattern deformation amount. This allowable pattern deformation amount can be set for each attribute of the wiring, and can be made variable depending on where the pattern is complicated and when it is not.

Curve (solid line in FIG. 50) forming the outer shape of the exposure pattern obtained by the lithography simulator as the reference pattern
Using, it is possible to perform defect inspection while verifying the validity of the simulation. The output data of the litho simulator is a light intensity distribution obtained by optically simulating. The contour curve is obtained from this distribution. The allowable pattern deformation amount in this case sets an error allowed as a simulation.

In this embodiment, design data is used as a reference pattern.

FIG. 6 is a flowchart showing an example of a recipe registration process in this embodiment. First, the operator inputs design data search parameters (here, parameters for designating the type of a sample to be inspected and a process), an inspection mode, an image acquisition parameter (an inspection area) to the reference pattern generation unit 11 via the input device 4. , Image size, number of pixels, slot number for specifying a wafer, and adjustment parameters of an optical system), and parameters for edge detection and inspection (step S202).

The following information is set as parameters for edge detection and inspection.

(R1) The amount of pattern deformation to be obtained (R2) The-side limit and the + side limit of the allowable pattern deformation amount, and the limit of the allowable edge difference in the edge direction (R3) Edge detection parameters empirically determined from the image quality (R4) Extraction rules for automatically recognizing pattern attributes (corners, straight lines, end points, isolated patterns, etc.) (R5) Length of profile acquisition section, interval between profile acquisition sections, profile acquisition section, profile (R6) The flag reference pattern generation unit 11 determines whether the profile acquisition section is variable and the measurement is determined at the time of measurement. The key database 21 is searched using the design data search parameters (product type and process of the sample to be inspected) as a key, and the design data is retrieved. Ri put out (step S204). The backbone database 21 is a database that stores design data (CAD data) for a pattern image to be inspected.

Next, the reference pattern generator 11 generates a reference pattern based on the design data (step S20).
6).

When it is necessary to perform shrink processing (processing for changing magnification), size processing (processing for changing line width), etc. on the design data so as to be most suitable for the position of the edge detected from the pattern image to be inspected. There is. Since the positions of edges to be detected are generally different between the first edge detection and the second edge detection, if necessary, two types of reference patterns are used for the first edge detection and the second edge detection. prepare.

Since the inspection is performed for each inspection unit area obtained by dividing the input inspection area by the image size, the reference pattern is also generated accordingly. The tests include sequential tests and random tests.

FIG. 8 is a diagram for explaining the sequential inspection. The inspection area is usually determined as a sum of rectangles. That is, the inspection area is not set as a unit for the entire surface of the wafer, but is set as a plurality of areas designated by rectangles (such as an upper short rectangle and a lower long rectangle as shown in FIG. 8). In order to inspect a region at high speed, a sequential scan is performed for each inspection unit region. A reference pattern is created for each inspection unit area.

FIG. 9 is a diagram for explaining the random inspection. In the random inspection, a certain region is not inspected sequentially but inspected pinpointly. In FIG. 9, only the inspection unit areas 301 to 304 are inspected.

FIG. 10 is a diagram showing an example of a reference pattern, and FIG. 11 is a diagram showing an example in which the reference pattern of FIG. 10 is converted into an edge vector for each pixel. In FIG. 10, the reference pattern (dotted line) is shown with sub-pixel accuracy. Usually, the edge direction of the reference pattern is parallel to the horizontal direction (x direction) or the vertical direction (y direction) of the pixel. The edge of the reference pattern also has information of the starting point (sub-pixel accuracy), direction, and amplitude for each pixel, similarly to the edge of the pattern image to-be-inspected. In this embodiment, the amplitudes of the edges of the reference pattern are all 1
I have to.

As shown in FIG. 12, the reference pattern may include a curve. To convert a reference pattern including a curve into an edge vector, for example, a pixel center 2
Tangent line 26 at point 262 on the reference pattern closest to 61
3 is an edge vector.

Next, the reference pattern generation unit 11 stores the reference pattern, the type of the sample to be inspected, the process, the inspection mode, the image acquisition parameters, and the parameters for edge detection and inspection in the recipe database 2.
2 (step S208). These data are called a recipe, which is a set of inspection parameters, and are managed using a product type, a process, and an inspection mode as keys.

(Inspection Processing) FIG. 13 is a flowchart showing an example of the inspection processing in this embodiment. First, the operator inputs recipe search parameters (here, product type, process, and inspection mode) to the inspection unit 12 via the input device 4 (step S302).

The inspection unit 12 searches the recipe database 22 using the recipe inspection parameters as keys, and retrieves the recipe (step S304). Then, in order to acquire the inspection target pattern image (optical image, electron beam image, focus ion beam image, probe microscope image, etc.),
Instruct the image generation device 7 to obtain image acquisition parameters,
An instruction is given for slot conveyance, alignment, and adjustment of the optical system (step S306). What is alignment?
This is a function for calculating a conversion coefficient between a coordinate system used by the D data and a coordinate value for managing an actual wafer observation position. This is embodied in CAD navigation. In CAD navigation, after alignment, coordinate values to be observed on the CAD data are converted into coordinate values for managing the observation position of the actual wafer, and the field of view of the imaging device is moved to that position to obtain an image at that position. The method is well known.

As the image generating device 7, a commonly used wafer defect inspection device, a CD-SEM or various microscopes can be used.

The image generation device 7 performs the following for each inspection unit area.
The inspection target pattern image (and its center position) is output to the inspection unit 12 (Step S308).

(First Edge Detection) Next, the inspection unit 12
Performs the first edge detection from the inspection target pattern image (step S310). For example, there are the following two methods for edge detection.

(A) One method is suitable when there is a contrast between the inside of the pattern and the background. Most of such images can detect edges by binarization processing,
If the contrast is relatively unclear, edges cannot be clearly detected. At this time, for example, [Reference 1]: R.
M. Haralick, “Digital step edges from ZERO crossin
g of second directional derivatives ”, IEEE Trans.
Pattern Anal. Machine Intell., Vol. PAMI-6, No.1, p
Edges can be obtained by applying the method disclosed in p.58-68, 1984. According to this method, an inflection point at an edge portion can be obtained with an accuracy of about 1/10 of a pixel unit.

(B) Another is that only the edges are bright,
A method that can cope with the case where there is no contrast between the inside of the pattern and the background. For example, [Reference 2]: “Cartan Stege
r. An unbiased detector of curvilinear structure
s ”, IEEE Trans. Pattern Anal. Machine Intell., 20
(2), The edge is obtained by the method disclosed in February 1998. According to this method, the peak of the edge portion can be obtained with an accuracy of about 1/10 of a pixel unit. However, in this method, the direction of the edge has only a value of 0 to 180 degrees. That is, the inside of the pattern cannot be specified.

An edge obtained by the above method using an edge amplitude image obtained by applying a differential filter (for example, a Sobel filter or a band-pass filter) to an image having a contrast between the inside of the pattern and the background. Good. In this case, the inside of the pattern can be determined, and the direction of the edge can be specified.

Since these methods use a window that is large to some extent, not only accuracy of about 1/10 of a pixel unit is obtained, but also the direction of the edge is stable. This means that it is not always necessary to connect edges and perform linear approximation.

In the edge detection in step S310, the amplitude and direction of the edge are determined in pixel units from the pattern image to be inspected. The amplitude takes a larger value as the edge becomes clearer. (A) In the case of an image having a contrast between the inside of the pattern and the background, the absolute value of the primary differential value of the image is used as the amplitude and the secondary differential value of the image Can be used as the edge position. On the other hand, in the case of (B) an image in which only edges are bright, for example, using the method of the above-mentioned document 2, the sign-inverted value (absolute value) of the second derivative of the image is set as the amplitude, Can be used as the edge position. In each case, the edges are obtained with sub-pixel accuracy. In the case of the image (A), a direction from 0 to 360 degrees can be defined. In the case of the image (B), the direction can be defined from 0 to 1 degree.
Only directions up to 80 degrees can be defined. This is due to the fact that the inside of the pattern cannot be specified from the local information in the image of FIG.

FIG. 14A is a diagram showing an example of an image (pattern image to be inspected) having a contrast between the inside of the pattern and the background, and FIG. 15 is a diagram showing edges detected from the image of FIG. is there. FIG. 14 shows the luminance value for each pixel. As shown in FIG. 15, an edge is detected for each pixel, and information of a starting point (sub-pixel accuracy), a direction (0 to 360 degrees), and an amplitude is obtained for each pixel. As described above, the amplitude takes a larger value as the edge becomes clearer.

FIG. 16 is a diagram showing an example of an image (pattern image to be inspected) in which only the edges (B) are bright.
FIG. 8 is a diagram showing an edge detected from the image of FIG. 7A. FIG. 16 also shows the luminance value for each pixel. Further, as shown in FIG. 17, an edge is detected for each pixel, and information of a start point (sub-pixel accuracy), a direction (0 to 180 degrees), and an amplitude is obtained for each pixel.

(Matching) Next, the inspection unit 12 obtains an expanded edge by expanding the edge of the pattern image to be inspected (step S312). In the present embodiment,
It is expanded by an allowable pattern deformation amount that is allowed in terms of electrical characteristics. At this stage, the allowable pattern deformation amount is a positive integer. This value is a value obtained by converting the larger of the absolute values of the negative limit and the positive limit of the (R2) allowable pattern deformation amount into an integer. By expanding by the allowable pattern deformation amount, matching can be performed while permitting a shape difference within an electrically allowable range.

FIG. 18 is a diagram showing an example of the edge amplitude of a one-dimensional pattern image to be inspected, and FIG. 19 is a diagram showing an example in which the edge in FIG. 18 is expanded. 18 and 19
In order to simplify the description, an example is shown in which the edge amplitude of each pixel has a scalar value in one dimension. To treat the deformation within the allowable pattern deformation amount equally,
A maximum value filter having a window twice as large as the allowable pattern deformation amount is applied. The maximum value filter obtains the maximum value of each pixel in a window that is in the vicinity of the target pixel, and uses that value as the value of the filtered pixel. In FIG. 19, the edge of FIG. 18 is expanded left and right by two pixels. This means that the allowable pattern deformation amount is 2
This is an example in the case of pixels.

On the other hand, it is assumed that the edge of the reference pattern is as shown in FIG. When the evaluation value (degree) of the matching is obtained from FIG. 19 and FIG. 20, the evaluation value of the matching is the same even if the pattern image to be inspected is shifted one pixel or two pixels to the left or right even at the current position. Become.

In order to avoid this, as shown in FIG. 21, expansion may be performed with weighting. This means that the smaller the allowable pattern deformation amount, the better. To achieve the expansion of FIG. 21, 0.5, 0.75,
A smoothing filter of 1.0, 0.75, or 0.5 may be used. In the case of FIG. 21, the evaluation value decreases if the pattern image to be inspected is shifted even by one pixel to the left or right.

Here, it is assumed that the edge of the reference pattern is wider by two pixels as shown in FIG. When the evaluation value is obtained from FIG. 21 and FIG. 22, the same evaluation value is obtained even if the inspection target pattern image is shifted by one pixel to the left or right even at the current position.

In order to avoid this, weighting may be performed as shown in FIG. In order to realize the expansion shown in FIG. 23, a 0.5, 0.9, 1.0, 0.9, 0.5 smoothing filter (FIG. 24) may be used. The coefficients of the smoothing filter should be determined experimentally.

From the above, the expansion as shown in FIG. 23 is desirable, but the expansion as shown in FIGS. 19 and 21 can also be used from the viewpoint of the processing speed and the degree of edge embedding.

FIG. 25 is a diagram showing an example of the amplitude of the edge of the two-dimensional pattern image to be inspected.
FIG. 26 is a diagram showing an example in which the edge of FIG. 25 is expanded. FIG.
At 5, the amplitude values are all 0 except at 20. FIG. 26 shows the result when the same expansion as in FIG. 19 is performed, and FIG. 27 shows the result when the same expansion as in FIG. 23 is performed.

FIG. 28 is a diagram showing an example of an edge vector of a two-dimensional pattern image to be inspected.
0 is a diagram showing an example in which the edge vector in FIG. 28 is expanded. FIG. 29 shows the result when the same expansion is performed as in FIG. 19, and FIG. 30 shows the result when the same expansion is performed as in FIG. The expansion is performed for each of the x and y components.

The inspection section 12 compares the dilated edge (the edge obtained by dilating the edge of the pattern image to be inspected) with the edge of the reference pattern, and performs matching on a pixel basis between the pattern image to be inspected and the reference pattern. (Step S314).

In the present embodiment, as will be described later, since matching is performed with sub-pixel accuracy, matching is performed in pixel units for the purpose of speeding up. Therefore, FIG. 31 shows FIG. 11 in units of pixels.

[0096] In matching in this embodiment, by shifting vertically and horizontally the reference pattern in pixels relative to the inspection object pattern image, the evaluation value F ₀ is the matching position the position of maximum (Figure 32). In this embodiment, as it follows, and the amplitude sum of the dilated edges in the pixel where the edge of the reference pattern exists as the evaluation value F _0.

[0097]

(Equation 1)

Here, E (x, y) is an edge vector having the amplitude of the expanded edge as its magnitude and the direction of the expanded edge as its direction. In a place where no edge exists, the magnitude of E (x, y) is zero. R (x + xs, y + ys) is an edge vector having the edge direction of the reference pattern as its direction. However, the magnitude of R (x + xs, y + ys) is 1 at a location where an edge exists, and is 0 at a location where no edge exists. Here, (xs, ys) is a shift quantity S ₁ of the edge of the reference pattern.

If only pixels for which R (x, y) is not 0 are stored in the calculation of F _0, the calculation can be performed at high speed and the storage area is small. Residual sequential test method (SSDA: Sequential Similarity Detect) using the sum of pixel amplitude values as an evaluation function
The calculation speed is further increased by using the high-speed calculation truncation used in the ion algorithm.

FIGS. 33 and 34 are diagrams in which FIGS. 29 and 31 are superimposed. In FIG. 33, pixel 2
Reference numeral 54 corresponds to the pixel 251 in FIG. 29 and the pixel 252 in FIG. In FIG. 34, the pattern image to be inspected is shifted one pixel to the right and one pixel downward from the state of FIG. Thus, pixel 255 corresponds to pixel 251 in FIG. 29 and pixel 253 in FIG. When using the evaluation value F ₀ is, the larger the degree of pixels present in the edge overlap, evaluation value increases. When using the evaluation value F ₀ may be performed dilation processing as shown in FIGS. 25 27. The evaluation value F ₀ can be applied to any of the images (A) and (B).

[0102] In the present embodiment uses the evaluation value F _0, it is also possible to use other evaluation values. For example, in the case of an image having a contrast between the (A) inside of the pattern and the ground it is considered to use the following evaluation value F _a.

[0102]

(Equation 2)

[0103] Also, for example, in the case of (B) only the edge is bright image, it is considered to use the following evaluation value F _b.

[0104]

(Equation 3)

[0105] When using the evaluation value F _a or F _b is 2
The expansion processing shown in FIGS. 8 to 30 may be performed. However, when performing the expansion as shown in FIG. 29, the expansion is performed for both the maximum value of the positive value and the maximum value of the negative value,
Select the one that takes a larger value in the calculation of the inner product.

[0106] When comparing the evaluation value F ₀ and an evaluation value F _a and F _b, the evaluation value F ₀ is directed to the high-speed processing for data is a scalar. On the other hand, the evaluation values _Fa and _Fb are effective, for example, in the case as shown in FIG. That is, the evaluation values _Fa and _Fb
Is used, the edge vector of the vertical line portion of the reference pattern (FIG. 35A) and the inspection target pattern image (FIG. 35A) are used.
The inner product of the horizontal line portion of (b)) and the edge vector is 0
Therefore, the portion 101 and the portion 102 are well matched. However, when the evaluation value F _0, the direction is to determine only the amplitude regardless there is a possibility that the 101 parts and 103 parts of thus matching.

Next, when comparing the evaluation value F _a and F _b, for example, as shown in FIG. 36, the interval spacing of the wiring 111, 113 and the base 112, 114 used F _a if the same both results preferable to F _b is obtained so can be distinguished whether the line.

In the present embodiment, the matching is performed by expanding the edges of the pattern image to be inspected. Alternatively, the matching may be performed by expanding the edges of the reference pattern.

The weighting can be changed depending on the position of the edge of the reference pattern. This is performed in the following procedure.

In FIG. 37, (a) shows an example of a reference pattern, and (b) shows an example of a reference pattern (solid line) and an image of a pattern to be inspected (dotted line) in (a). The reference pattern shown in FIG. 37A is a periodic pattern, but has one gap. When matching the reference pattern with the pattern image to be inspected, as shown in FIG. 37 (b), even if both patterns are displaced, they match except for the gap, so the evaluation value of the matching is Will be expensive. Therefore, it is conceivable to increase the weight of the gap portion so that the matching evaluation value is greatly reduced when the gap of the inspection target pattern image does not match the gap of the reference pattern.

As a procedure for weighting, first, the period of the pattern is obtained by the autocorrelation method. Next, the original pattern is compared with the pattern shifted by one cycle, and the pattern that is not in the pattern shifted by one cycle is obtained. And it recognizes the obtained pattern as a unique pattern,
The degree of contributing to matching (weighting) is made stronger than other patterns. An empirical value (1 or more) is used for the amplitude of the reference pattern to express the degree of contribution. As this value, a fixed value or a ratio of fixed value / unique pattern in all patterns is effective.

When the matching is performed and the shift amount S ₁ = (xs, ys) at which the maximum evaluation value is obtained is obtained, the reference pattern is shifted by S ₁ . Subsequent processing is performed with this shift performed.

The shift amount S ₁ can be output to the display device 5 and the printing device 6 as the inspection result.

After the matching is completed, the pattern image to be inspected is binarized. The binarization is performed by judging the presence / absence of the edge amplitude using one of edge detection parameters (threshold) in the recipe. As a method of binarization, the number of pixels × p (normally about 0.9 to 1.1) corresponding to the edge of the reference pattern becomes 1,
There is also a method of binarizing the edge image of the inspection target pattern image (p tile method). The above threshold value or p may be set as the parameter of (R3).

(First Inspection) Next, the inspection unit 12
Inspection. Specifically, calculation of the pattern deformation amount and defect detection are performed.

The inspection unit 12 first associates the edge of the pattern image to be inspected with the edge of the reference pattern (step S318).

The position of the edge is handled with sub-pixel accuracy. Therefore, the distance between edges is also obtained with sub-pixel accuracy. The direction is determined as a value of 0 to 360 degrees, for example, with the right direction being 0 degrees.

For each edge pixel of the reference pattern, an edge pixel of the pattern image to-be-inspected that is within the distance of the allowable pattern deformation amount corresponding to (R2) is searched. Then, among the detected edges, those whose direction difference from the edge of the reference pattern is equal to or smaller than the allowable direction difference of the edge of (R2) are associated as edges within the allowable deformation. That is, in this embodiment, the correspondence is performed in consideration of the distance between the edge of the pattern image to be inspected that has been matched and the edge of the reference pattern, and the directions of both edges. The vector d (x, y) between the associated two edges can be used to determine the pattern deformation amount.

When there are a plurality of candidates for association, the candidate having a small distance and a small direction difference is preferentially associated.

FIG. 38 is a diagram showing an example of correspondence between the edge of the pattern image to be inspected and the edge of the reference pattern. In FIG. 38, edges are indicated by arrows to indicate the direction. In the example of FIG. 38, in each pixel including the edge of the reference pattern, the association is performed by searching for the edge of the inspection target pattern image in the direction perpendicular to the edge direction from the center of the edge of the reference pattern. . If an edge of the pattern image to-be-inspected whose distance is within the allowable pattern deformation amount and the direction difference is equal to or smaller than the allowable direction difference of the edge is found, the two edges are associated with each other.
In FIG. 38, the vector d between the associated two edges
(X, y) is shown for reference.

In FIG. 39, (a) shows an example of the edge of the reference pattern, and (b) shows an example of the edge of the inspection target pattern image corresponding to the reference pattern of (a). An example of associating both edges will be described with reference to FIG. In this example,
The allowable pattern deformation amount is one pixel. The allowable direction difference of the edge is 60 degrees. For example, when the edge of the pattern image to-be-inspected corresponding to the edge 81 of the reference pattern is searched, the edge 68 is within the distance of the allowable pattern deformation amount of the edge 81, and the difference in the direction is equal to or smaller than the allowable direction difference of the edge. Because of this, it is recognized as an edge corresponding to the edge 81. With respect to the edge 84 of the reference pattern, the edge 70 is recognized as the corresponding edge of the pattern image to be inspected. At this time, with respect to the edge 82 of the reference pattern, the edge 61 is not within the distance of the allowable pattern deformation amount. The edge 64 is not within the distance of the allowable pattern deformation amount, and the direction difference is larger than the allowable direction difference of the edge. Although the edges 66 and 69 are within the distance of the allowable pattern deformation amount, the direction difference is larger than the allowable direction difference of the edge. Therefore, no edge corresponding to edge 82 is found.
Similarly, the edge 83 cannot be found.

Although the example of FIG. 39 is a method that does not distinguish between the inside and the outside of the pattern and has only a value of 0 to 180 degrees in the direction, a method of distinguishing between the inside and the outside of the pattern is also possible. is there. For example, if the edge direction is determined so that the inside of the pattern is always on the right hand, FIG.
Is as shown in FIG. 40, and the association can be performed more strictly.

Next, the inspection section 12 detects a defect (step S320). If a defect is detected, the defect information (here, defect position, size information, and image) is output to the defect type recognition unit 14 (steps S322 and S322).
4).

The defect type recognizing unit 14 determines the defect type based on the defect information and the information in the defect type reference database 23 (step S 326), that is, obtains the feature amount from the given image and obtains the defect type reference image database The defect type is determined by collating with the feature amount of the image stored in the. The defect type recognition unit 14 outputs the defect information and the defect type to the display device 5 and the printing device 6 via the output unit 13 (Step S328). Here, the defect type reference database 23
Is an image obtained by registering an already acquired image for each defect type.

As a method of recognizing a defective area, a method of recognizing an area from an edge of an image of a pattern to-be-inspected that cannot be associated and recognizing the area as a defective area (recognition method A) can be considered. This is effective for detecting a defect having a clear edge. However, since it is weak against the detection of a defect whose edge is unclear, in such a case, an area is recognized from the edge of the associated pattern image to-be-inspected, and the distribution of pixel luminance values in that area is non-uniform. Method for recognizing similar parts as defective areas (recognition method B)
Is suitable. That is, a defect is recognized from an abnormality in the luminance value distribution.

In the recognition method A, the pixels of the edges (eg, edges 61 to 67, 69, and 75 in FIG. 39A) of the pattern image to-be-inspected that cannot be associated are recognized as defects. The inspection unit 12 dilates these pixels and connects the pixels. As a process for expanding such a bitmap (binary image), a process called morphology is known. Next, the pixels connected by the labeling process are recognized as a lump as one area. Here, the labeling process is a method of writing the same value to pixels connected in the vicinity of 4 or 8 and generating a connected pixel group. By giving a different value to pixels that are not connected, the connected pixel group can be distinguished. The foreign matter is defined as a unit that can be separated as the mass region, and its outer shape is recognized. Paint the pixels inside from the outline by the paint process. With these pixels as defects, the center of gravity and size of the defects are calculated.

In the recognition method B, the edges of the pattern images to be inspected which have been associated are connected to form a region. In each of the inner and outer regions, a portion excluding a boundary (edge) is obtained as a cluster of pixels. With respect to the inner and outer regions of the cluster of pixels, a pixel luminance value is obtained from the inspection target pattern image obtained first. These values can be expected to form a normal distribution if there are no defects. That is, defective pixels can be detected by applying a quality control method. In a normal case, there should be little change in luminance in each of the inner region and the outer region. Therefore, it is possible to detect an area in which the distribution of luminance is non-uniform among the areas in the inspection target pattern image and recognize the area as a defective area. The obtained defective pixels are recognized as a mass, and the center of gravity and the size are calculated.

FIG. 41 is a diagram showing an example of a pattern image to be inspected. A broken line 201 indicates an edge of the inspection target pattern image. Solid lines 202 and 203 on both sides of the broken line 201
Is a line segment whose edge is widened by a specified width.
The part surrounded by 203 is recognized as an edge area. Ground 2
The luminance values of the pattern 04 and the inside 205 of the pattern substantially form a normal distribution.

As shown in FIG. 42, the portion D exceeding about ± 3σ is highly likely to be a foreign matter. D includes noise, but noise exists relatively uniformly in the region,
Foreign matter is present as a solid. A binarized map is created in which pixels having a luminance value of D are 1 and pixels having other luminance values are 0. A block of pixels with a 1 less than or equal to a specified size (eg, 2 × 2 pixels) (eg,
The block of pixels 207) in FIG. 41 is deleted. A median filter or morphology filter can be used.
This size is an empirical value in consideration of the size of the foreign substance to be detected. A cluster of pixels with a remaining one (for example,
The cluster of pixels 206) in FIG. 41 is regarded as a foreign substance.

The defect type recognition unit 14 can perform automatic defect type classification as follows. That is, a geometric feature of a cluster of pixels recognized as a defect is obtained. As a result, it is possible to grasp the shape characteristics of the defect such as round or long, and it is possible to determine that the defect is a foreign substance if the defect is long and that the defect is a scratch if it is long. Pixels recognized as defects are divided into three parts: inside, outside, and boundary of the pattern. For each of these parts, a feature amount using the pixel luminance value of the pattern image to-be-inspected first obtained is obtained. For example, when it is determined that the foreign matter is a foreign matter based on the geometrical feature amount, it is possible to determine whether the foreign matter is a metal piece or an organic matter (for example, human red) based on the obtained feature amount. That is, if the foreign matter is a metal, the type can be determined as being bright because the reflection is strong, and if the foreign matter is dark, it is dark. Further, when the luminance of a pixel recognized as a foreign substance inside the pattern has a large fluctuation, it is determined that there is a high possibility that the foreign substance exists on the pattern, and conversely, the fluctuation of the luminance is small. In this case, it is determined that there is a high possibility that the foreign matter exists below the pattern. This is a difficult process with the conventional die-to-die method. Using these feature amounts, the defect type is determined by a well-known classification method. As a classification method, a method of performing comparison with a defect type reference image database by using the k shortest distance method and determining is effective.

Such automatic classification of defect types is performed by a conventional optical or SEM ADC (Automatic Defefe).
ct Classification), but according to the method of the present invention using design data, the distinction between the inside and the outside of the pattern can be clearly distinguished, so that the feature amount of each part can be accurately grasped, and the classification accuracy can be improved. improves.

Next, the inspection section 12 obtains the amount of pattern deformation from the relationship between the edge of the associated inspection target pattern image and the edge of the reference pattern (step S33).
0). The pattern deformation amount is obtained for a portion where no defect is detected as a result of the defect detection. Then, the display device 5 and the printing device 6 determine the amount of pattern deformation via the output unit 13.
(Step S332).

As the pattern deformation amount, a pattern deformation amount obtained from the entire image and a pattern deformation amount obtained for each pattern attribute can be considered.

As the pattern deformation amount obtained from the entire image, for example, a positional shift amount, a magnification change amount, and a line width thickening amount can be considered.

The amount of displacement is obtained as the average value of the vector d (x, y) between the associated edges. This is the shift amount (correction amount) S ₂ with sub-pixel accuracy of S ₁ = (xs, ys). By correcting amount shifting the reference pattern is shifted with matching in pixels on the basis of the shift amount S _2, it is possible to perform matching with sub-pixel accuracy.

In order to obtain the amount of magnification change in the x direction, a regression line is obtained by approximating the x component of the vector d (x, y) relating to the reference pattern in the vertical direction with a regression line D (x). Then, the gradient of the regression line is set as the magnification change amount in the x direction. The same applies to the magnification change amount in the y direction.

In FIG. 43, (a) shows an example of the edge of the reference pattern (broken line) and the edge of the image of the pattern to be inspected (solid line), and (b) shows y = y between the edges shown in (a).
Vector d (x, y ₀₎ in the ₀ to x components of the regression line D
An example approximated by (x) is shown. When the x component of the vector d (x, y ₀ ) is approximated by a regression line D (x) = ax + b, the gradient a corresponds to the magnification change amount. In the example of FIG. 43A, it can be seen that the pattern of the inspection target pattern image is entirely larger than the reference pattern.

In FIG. 44, (a) shows another example of the edge of the reference pattern (broken line) and the edge of the image of the pattern to be inspected (solid line), and (b) shows y between the edges shown in (a).
An example is shown in which the x component of the vector d (x, y ₀ ) at = y ₀ is approximated by a regression line D (x). In the example of FIG.
In addition to the pattern of the inspection target pattern image being larger than the reference pattern as a whole, the width of the line (line) is large. In FIG. 44 (a), reference pattern lines (wirings) 121, 122, and 123 respectively correspond to lines 124, 125, and 126 of the pattern of the pattern image to-be-inspected.

The width of the line width in the x direction is, for example, sig
n (x, y ₀ ) · {x component of d (x, y ₀ ) −D (x)}
Can be obtained as an average value of Where sign
(X, y ₀ ) takes −1 if the position of (x, y ₀ ) is at the left end of the line, and takes 1 if it is at the right end of the line. In addition, regarding the amount of increase in line width, sign (x, y ₀ ) ·
If the variance of {x component of d (x, y ₀ ) −D (x)} is obtained, a line width variation index can be obtained.

Next, the pattern deformation amount obtained for each pattern attribute will be described. As attributes of the pattern, a corner 171, a long wiring 172, a tip 173, an isolated pattern 174, and the like can be considered (FIG. 45). As the pattern deformation amount related to the attribute of the pattern, for example, in addition to the above-described displacement amount, magnification change amount, and line width increase amount, the deformation amount of a feature amount such as area, perimeter, circularity, moment, radius of curvature, etc. Can be considered.

The pattern attribute can be automatically added to the reference pattern. However, attributes can be added manually. Add (extract) pattern attributes
The rule for this is set as (R4) when the recipe is created.

FIG. 46 is a diagram showing the displacement of the tip. As shown in FIG. 46A, the displacement of the leading end is a distance from the edge 164 of the reference pattern to the edge 163 of the pattern image to be inspected (in the direction perpendicular to the edge of the reference pattern). For example, a distance between a point closest to the edge 164 of the reference pattern in the edge 163 of the pattern image to be inspected and the edge 164 of the reference pattern can be measured as the positional deviation amount of the leading end.

Further, as shown in FIG. 46 (b), the average value, maximum value, minimum value, median value, standard deviation, etc. of a plurality of distances measured for a section 157 having an arbitrary width are determined by the positional deviation of the tip. It may be an amount.

In FIG. 46, the positional deviation amount at the tip has been described. However, the positional deviation amount can be similarly measured for a long wiring, a corner, a portion where an attribute is connected to an attribute, and the like. Further, for example, for a corner, a positional shift amount in a direction having a half angle formed by the corner or a direction having a designated angle can be measured.

FIG. 47 is a diagram showing the amount of displacement of the center of gravity of an isolated pattern. The amount of displacement of the center of gravity is calculated based on the center of gravity 16 of the edge 160 of the reference pattern (constituting an isolated pattern).
2 and the amount of displacement between the center of gravity 161 of the edge 159 of the pattern image to be inspected (which constitutes an isolated pattern).

In FIG. 47, it is conceivable to measure the amount of deformation of the feature amount (area, perimeter, circularity, moment, etc.) of the isolated pattern. That is, it is conceivable to measure a difference between the feature amount of the edge 160 of the reference pattern and the feature amount of the edge 159 of the pattern image to-be-inspected.

In FIG. 48, (a) shows an example of the corner of the edge of the reference pattern, and (b) shows an example of the corner of the edge of the image of the pattern to-be-inspected. The corner of the edge 166 of the reference pattern shown in FIG. 48 (a) is rounded. As the radius of curvature of the corner, for example, a major axis, a minor axis, or a radius obtained by least-squares approximation of the curve of the corner with an ellipse or a circle can be used. By determining the radius of curvature of the corner of the edge 166 of the reference pattern and the radius of curvature of the corner of the edge 165 of the pattern image to-be-inspected, the amount of deformation of the radius of curvature of the corner can be determined.

The above inspections can be applied simultaneously (with one imaging) to a plurality of locations within one imaging range (within the field of view) instead of performing each inspection one by one.

The inspection items are selected in accordance with the above-mentioned recipe item (R1) the pattern deformation amount to be obtained.

The pattern attribute extraction rule ((R
There are various types 4)), examples of which will be described with reference to FIG. The corner is extracted as a vicinity of a contact point of two lines that contact at a predetermined angle (90 degrees, 270 degrees, or the like). The long wires are separated as two parallel line segments having a line width apart and having a length equal to or longer than the specified length. The tip is a line segment having a line width length, and other line segments of a predetermined length or more are provided at both ends thereof.
It is extracted as a part that contacts at an angle of 0 degrees.
The isolated pattern is extracted as a closed figure having a predetermined area or less.

(Second Edge Detection) The inspection unit 12 detects an edge again from the pattern image to be inspected for a portion where no defect is detected as a result of the defect detection (step S334).

The edge detection of the inspection target pattern image is performed by obtaining a profile on the inspection target pattern image based on the second reference pattern. Here, the second
As the reference pattern, a reference pattern in the case where the position of the point Q in FIG. 53 is considered as an edge is used. On the other hand, as the first reference pattern, for example, in the case of the above-described (B) image in which only the edge is bright, a reference pattern in a case where the position of the point P is considered as an edge is used. Therefore, the second
Is generally different from the first reference pattern.

Before detecting the edge of the pattern image to be inspected, the second reference pattern is shifted by the shift amount S ₁ + S ₂ . Subsequent processing is performed with this shift performed.

Various methods, such as a threshold value method and a linear approximation method, are disclosed for obtaining the edge position from the profile. In the present embodiment, the threshold value method is used to determine the C position.
The line width measurement performed by the D-SEM is applied to a two-dimensional pattern (pattern image to be inspected). However, the same processing can be performed by replacing the threshold value method with another method such as a linear approximation method. Here, the straight line approximation method is a method of approximating a profile with a straight line and specifying an edge position using an intersection.

For edge detection, two methods are conceivable. One is a method in which a direction and a position for taking a profile are set in advance for a second reference pattern.

In the present embodiment, when a profile acquisition section is set in advance, it is performed at the time of recipe creation as described above. In this case, the (R6) profile acquisition section of the above-described recipe item is variably set, and the flag of whether to determine at the time of measurement is off, and the profile acquisition section is set in advance with respect to the second reference pattern. become.

The section for acquiring the profile (profile acquisition section) is the length of the profile acquisition section of (R5),
For example, based on the interval between the profile acquisition section and the profile acquisition section, as shown in FIG.
Are set in the vertical direction of the second reference pattern with the reference pattern of (1) as the middle point (double line in the figure). The second shown in FIG.
As described with reference to FIG. 7, the reference pattern is corrected so that the corners are rounded.
Further, instead of the above-mentioned second reference pattern, as shown in FIG. 50, it is also possible to use a curve (solid line in the figure) which forms the outer shape of the exposure pattern obtained by the litho simulator.

In the second edge detection processing, at the position (section) corresponding to the above-mentioned profile section in the pattern image to-be-inspected, the profile is determined based on the interval for examining the luminance value within the (R5) profile acquisition section. create. The interval is usually an arbitrary value equal to or less than the pixel interval, and the length of the profile section is an arbitrary length longer than the allowable deformation amount of the pattern. The profile is created using a technique such as bilinear interpolation, spline interpolation, or Fourier series.

FIG. 51 is an enlarged view of a part (part B) of FIG. 49, and FIG. 52 is an enlarged view of a part (part C) of FIG. The double line in the figure is the profile acquisition section, the intersection of the grid indicates the position of the pixel, and the black point indicates the position where the luminance value of the pattern image to be inspected is examined.

The bilinear interpolation method is (0,
0) (0,1) (1,0) (1,1), the luminance values I (0,0), I (0,1), I (1,
0), I (1, 1), the position (x, y), (0 <
The luminance value I (x, y) at the point where x ≦ 1, 0 <y ≦ 1) is calculated by the following formula.

[0161]

## EQU4 ## I (x, y) = [I (0, 0) (1-x) + I
(1,0) x] (1-y) + [I (0,1) (1-x)
+ I (1,1) x] y

From the profile thus obtained, the second edge position is determined by applying the threshold method. FIG.
As shown in FIG. 3, the maximum luminance value V and its position P in the obtained profile are obtained. A value obtained by multiplying the maximum luminance value V by a coefficient k specified in advance is set as a threshold value T, and an intersection between a straight line of luminance value = threshold value T and a profile curve is obtained. At these intersections, an intersection Q that is located outside of the pattern from the point P and is closest to the point P is obtained. This intersection Q is obtained for all the profiles and set as the second edge position.

The coefficient k plays a role in determining the second edge position. That is, since the cross-sectional shape of the actual wiring formed on the wafer has a trapezoidal shape, the edge position is controlled on the upper side, on the lower side, or on the middle of the trapezoid, and is adjusted by the coefficient k. be able to.

After the above-described edges are obtained, curve approximation (including polygon approximation) is performed based on them, and a second edge is obtained. The simplest method is to simply connect as a polygonal line. However, as a method to connect smoothly using the least squares method, for example, the following method can be used. That is, as shown in FIG. 54 (a), T. Pavlidis and SLHo
rowitz: “Segmentation of plane curves”, IEEE Tr
ans. on Computers, vol. C-23, no. 8 Aug., 1974. Alternatively, a curve approximation by smoothing plane data using a least squares method and a two-dimensional spline function as shown in FIG. 54B can be used. The former can be processed at high speed, but is inflexible for those containing many rounded shapes. On the other hand, the latter has the property of satisfying high-speed operation and having flexibility. In addition to these, various methods such as a method using a Fourier descriptor are disclosed, and these can be replaced.

The above curve approximation can be performed even after the first edge detection is performed.

Next, as another method different from the above method, there is a method in which the profile acquisition section is made variable and the profile acquisition section is determined at the time of edge detection. That is, as shown in FIG. 55 (a), this is a method of setting a profile acquisition section in the vertical direction of the first edge of the detected pattern to-be-inspected. According to this method, as shown in FIG. 55 (b), even if the first edge (solid line) of the inspection target pattern image is shifted from the second reference pattern (dotted line), the profile acquisition section is specified. , Edges can be detected.
This method is easier to follow the deformation of the pattern than the above method. After setting the profile acquisition section, the same processing as in the above method is performed.

The result of the second edge detection can be output to the display device 5 and the printing device 6.

The detected second edge is, for example, as shown in FIG.
Can be converted into an edge vector for each pixel by using the method described with reference to FIG. This edge vector corresponds to the edge vector obtained by the binarization processing obtained before the first inspection.

(Second Inspection) After the above-described second edge detection, the inspection unit 12 performs a second inspection (step S336).

This inspection is the same processing as the above-described first inspection.
Then, a defect is detected and the pattern deformation amount is obtained. This
The position shift amount (shift
Quantity) S _ThreeIs the above S_TwoCorresponding to S obtained here
_ThreeThe above-mentioned S₁And S_TwoIs the second standard
Between the pattern and the pattern of the pattern image to be inspected.
The shift amount.

The inspection result is output to the display device 5 and the printing device 6 via the output unit 13 (step S33).
8).

If the above processing has been performed for all the inspection unit areas, the inspection processing ends, otherwise, the flow returns to step S308 (step S340).

(Other Inspections) In the case of an SEM having a function of electromagnetically observing a part of a low-magnification image as a high-magnification image, it is possible to measure a pattern that cannot be covered by a high-magnification image. That is, it means that the edge position obtained in the high-magnification image can be accurately converted to the edge position obtained in the low-magnification image. The same relationship may be realized by a high-precision stage. For example, in FIG. 56, the positions 182 and 183 on the pattern 181 of the pattern image to be inspected are
84 and 185, the image is converted to a position on the low-magnification image 187, and the width 1 of the pattern 181 of the inspection target pattern image is obtained.
If 86 is obtained, the length can be measured more accurately than when the low-magnification image 187 alone is obtained.

(Inclination and Magnification Adjustment) In the inspection method described above, the inclination of the pattern image to be inspected can be adjusted before the inspection or at an appropriate time during the inspection, if necessary, using the pattern deformation amount technique. The magnification can be adjusted. That is, the inspection target pattern image and the reference pattern of the portion suitable for adjustment are acquired. By the affine transformation, several inspection target pattern images in which the inclination and the magnification that can be candidates are changed are obtained. The obtained inspection target pattern image is compared with the reference pattern, and the inspection target pattern image having the smallest pattern deformation amount is selected. The inclination and the magnification with respect to the selected pattern image to be inspected are registered as correction amounts. Instead of applying an affine transformation to the inspection target pattern image, a method of applying an affine transformation to the reference pattern may be used.

The affine transformation means a primary transformation using coefficients a to f.

[0176]

[Equation 5] X = ax + by + c Y = dx + ey + f

If information such as the amount of pattern deformation, the position and size of the defect area, the type of defect, and the statistics of the amount of pattern deformation and images obtained by the inspection method of the present invention is used, the circuit of the defect area can be used. , Analysis of the degree of influence on the circuit in the preceding and subsequent steps, and analysis of optimization parameters such as exposure conditions.

The pattern inspection apparatus according to the present embodiment can be considered as an apparatus that performs pattern matching if attention is paid to a portion that outputs a shift amount.

As described above, an example of the present invention has been described, but various other modifications are possible. For example, it is easy to transform acquired image data into an offline input processing method via an external input device such as a magneto-optical disk or a magnetic tape, or via a LAN such as Ethernet. It is also possible to adopt a hybrid method in which a typical die in a wafer is inspected by the method of the present invention, and then the other dies are inspected by die-to-die comparison.
Furthermore, the image generation method may be another method, and the design data is not limited to CAD, but may be another method. In the present embodiment, the inspection result and the like are output to the display device 5 and the printing device 6, but may be output to an image database, a simulator, a recording medium, or the like, or to another computer via a network. It may be transmitted (output).

[0180]

As described above, according to the present invention, a comparison inspection between a pattern image to be inspected and a reference pattern can be performed in real time.

Further, matching can be performed while allowing a shape difference within an electrically allowable range.

Further, stable defect detection can be performed.

Further, the measurement of the two-dimensional pattern (pattern image to be inspected) which has been performed visually so far can be quantitatively and automatically performed at high speed.

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a theoretical pattern based on design data.

FIG. 2 is a diagram showing an example of a pattern actually manufactured based on design data.

FIG. 3 is a diagram illustrating an outline of an inspection process performed by a pattern inspection apparatus according to the embodiment of the present invention.

FIG. 4 is a diagram illustrating a hardware configuration example of a pattern inspection apparatus according to the embodiment of the present invention.

FIG. 5 is a diagram showing a functional block diagram of a pattern inspection apparatus according to the embodiment of the present invention.

FIG. 6 is a flowchart illustrating an example of a recipe registration process according to the embodiment of the present invention.

FIG. 7 is a diagram illustrating an example of correction of a reference pattern.

FIG. 8 is a diagram for explaining a sequential inspection.

FIG. 9 is a diagram for explaining a random inspection.

FIG. 10 is a diagram showing an example of a reference pattern.

11 is a diagram illustrating an example in which the reference pattern of FIG. 10 is converted into an edge vector for each pixel.

FIG. 12 is a diagram illustrating an example in which a reference pattern including a curve is converted into an edge vector.

FIG. 13 is a flowchart illustrating an example of an inspection process according to the embodiment of the present invention.

FIG. 14 is a diagram showing an example of an image (pattern image to be inspected) in which contrast is provided between the inside of the pattern and the background.

FIG. 15 is a diagram showing edges detected from the image of FIG. 14;

FIG. 16 is a diagram illustrating an example of an image (pattern image to be inspected) in which only the contour is bright.

FIG. 17 is a diagram showing edges detected from the image of FIG. 16;

FIG. 18 is a diagram illustrating an example of an edge amplitude of a one-dimensional pattern image to-be-inspected;

19 is a diagram illustrating an example in which the edge in FIG. 18 is expanded.

FIG. 20 is a diagram illustrating an example of an amplitude of an edge of a one-dimensional reference pattern.

21 is a diagram showing another example in which the edge of FIG. 18 is expanded.

FIG. 22 is a diagram illustrating another example of the amplitude of the edge of the one-dimensional reference pattern.

FIG. 23 is a view showing another example in which the edge of FIG. 18 is expanded.

FIG. 24 is a diagram illustrating an example of a smoothing filter.

FIG. 25 is a diagram illustrating an example of an amplitude of an edge of a two-dimensional pattern image to-be-inspected;

26 is a diagram illustrating an example in which the edge in FIG. 25 is expanded.

FIG. 27 is a diagram showing another example in which the edge of FIG. 25 is expanded.

FIG. 28 is a diagram illustrating an example of an edge vector of a two-dimensional pattern image to-be-inspected;

29 is a diagram illustrating an example in which the edge vector in FIG. 28 is expanded.

30 is a diagram illustrating another example in which the edge vector in FIG. 28 is expanded.

31 is another diagram in which the reference pattern of FIG. 10 is represented by an edge vector in pixel units.

FIG. 32 is a diagram for explaining matching.

FIG. 33 is a view in which FIG. 29 and FIG. 31 are superimposed.

FIG. 34 is a view in which FIG. 29 and FIG. 31 are superimposed.

35A is a diagram illustrating an example of a reference pattern, and FIG. 35B is a diagram illustrating an example of a pattern image to be inspected.

FIG. 36 is a diagram showing an example in a case where a wiring interval is equal to a base interval.

37A is a diagram illustrating an example of a reference pattern, and FIG. 37B is a diagram illustrating an example of the relationship between the reference pattern of FIG.

FIG. 38 is a diagram illustrating an example of an edge of an inspection target pattern image and an edge of a reference pattern after matching is performed.

FIG. 39A shows an example of the edge of the reference pattern, and FIG.
FIG. 5 is a diagram illustrating an example of an edge of a pattern image to be inspected.

FIG. 40 is a diagram illustrating another example of a method for assigning direction information.

FIG. 41 is a diagram showing an example of a pattern image to be inspected.

FIG. 42 is a diagram illustrating an example of a distribution of frequencies with respect to luminance values.

43A shows an example of an edge of a reference pattern and an edge of an image of a pattern to-be-inspected, and FIG. 43B shows a vector d (x, y ₀ ) at y = y ₀ between the edges shown in FIG. FIG. 7 is a diagram showing an example in which the x component of the above is approximated by a regression line D (x).

44A shows another example of the edge of the reference pattern and the edge of the image of the pattern to-be-inspected, and FIG. 44B shows the vector d (x, y) at y = y ₀ between the edges shown in FIG. the x component of ₀₎ is a diagram showing an example approximated by the regression line D (x).

FIG. 45 is a diagram illustrating an example of a pattern attribute.

FIG. 46 is a diagram illustrating a positional shift amount of a tip.

FIG. 47 is a diagram showing the amount of displacement of the center of gravity of an isolated pattern.

FIG. 48A is a diagram illustrating an example of an edge corner of a reference pattern, and FIG. 48B is a diagram illustrating an example of an edge corner of an inspection target pattern image.

FIG. 49 is a diagram illustrating an example of a profile acquisition section.

FIG. 50 is a diagram showing a curve forming an outer shape of an exposure pattern obtained by a litho simulator.

FIG. 51 is an enlarged view of a part (part B) of FIG. 49;

FIG. 52 is an enlarged view of a part (part C) of FIG. 51;

FIG. 53 is a diagram showing an example of a profile.

FIG. 54 is a diagram illustrating an example in which curve approximation is performed based on a second edge position (point) to determine a second edge.

FIG. 55 (a) shows another example of the profile acquisition section, and FIG. 55 (b) shows the first edge and the second edge of the inspection target pattern image.
FIG. 6 is a diagram showing an example of a relationship with a reference pattern.

FIG. 56 is a diagram illustrating an example in which length measurement is performed using a high-magnification image and a low-magnification image.

[Explanation of symbols]

REFERENCE SIGNS LIST 1 Main control unit 2 Storage device 3 Input / output control unit 4 Input device 5 Display device 6 Printing device 7 Image generation device 11 Reference pattern generation unit 12 Inspection unit 13 Output unit 14 Defect type recognition unit 21 Basic database 22 Recipe database 23 Defect type Reference database 61 to 70, 75, 81 to 84 Edge 101 to 103 Part 111, 113 Wiring 112, 114, 204 Base 121 to 126 Line 151 Inspection target pattern image 152 Reference pattern 157 Section 159, 163, 165 Inspection target pattern image Edge 160, 164, 166 Reference pattern edge 161 Center of gravity of edge of inspection target pattern image 162 Center of gravity of edge of reference pattern 171 Corner 172 Long wiring 173 Tip 174 Isolated pattern 181 Pattern of inspection target pattern image 182, 183 Position 184, 185 High magnification image 186 Width of pattern of inspection target pattern image 187 Low magnification image 201 Dashed line 202, 203 Solid line 205 Inside pattern 206, 207 Pixel block 251 to 255 Pixel 261 Pixel 262 Pixel center 262 Pixel Point on reference pattern closest to center 263 Tangent line 301-304 Inspection unit area

Continued on the front page (51) Int.Cl. ⁷ Identification code FI Theme coat II (Reference) H01L 21/66 G01B 11/24 K 5L096 F F term (Reference) 2F065 AA12 AA56 BB02 CC17 DD06 FF04 PP24 QQ04 QQ18 QQ24 QQ27 QQ31 QQ38 RR08 SS06 SS13 2G051 AA51 AA56 AB02 AC21 EA11 EA12 EA14 EC01 ED15 2H095 BD01 BD28 4M106 AA01 BA02 BA03 BA04 CA39 DB05 DB21 DJ18 DJ20 DJ21 5B057 AA03 BA02 CA12 CA16 CB12 CB16 CE04 DA03 DB02 DC16 DC03 5

Claims (19)

[Claims] 1. A pattern inspection apparatus for inspecting an inspection target pattern by comparing the inspection target pattern with a reference pattern, comprising: a storage unit for storing the reference pattern; an input unit for inputting an image of the inspection target pattern; An inspection unit that inspects the inspection target pattern by comparing an edge of the image of the inspection target pattern and the edge of the stored reference pattern, and an output unit that outputs a result of the inspection. Characteristic pattern inspection equipment. 2. The pattern inspection apparatus according to claim 1, wherein the inspection means compares the edge of the pattern image to be inspected with the edge of the reference pattern, so that the pattern image to be inspected and the reference pattern are compared. A pattern inspection apparatus characterized by performing matching with a pattern. 3. The pattern inspection apparatus according to claim 2, wherein the matching is performed by expanding an edge of the pattern image to be inspected. 4. The pattern inspection apparatus according to claim 2, wherein the matching is performed by expanding an edge of the reference pattern. 5. The pattern inspection apparatus according to claim 2, wherein the matching is performed by summing a product of an edge amplitude of the inspection target pattern image and an edge amplitude of the reference pattern at each pixel. A pattern inspection apparatus characterized in that a pattern inspection is performed as an evaluation value. 6. The pattern inspection apparatus according to claim 2, wherein the matching is performed by summing an inner product of an edge vector of the pattern image to be inspected and an edge vector of the reference pattern at each pixel;
Alternatively, the pattern inspection apparatus is characterized in that a sum of absolute values of the inner products is used as an evaluation value, the edge vector has an edge amplitude as its magnitude, and an edge direction as its direction. 7. The pattern inspection apparatus according to claim 2, wherein the matching is performed by changing a weight for each part of the reference pattern. 8. The pattern inspection apparatus according to claim 1, wherein the inspection unit associates an edge of each pixel of the reference pattern with an edge of each pixel of the actual pattern image. Characteristic pattern inspection equipment. 9. The pattern inspection apparatus according to claim 8, wherein the association is performed by: a distance between an edge of each pixel of the reference pattern and an edge of each pixel of the pattern image to be inspected; A pattern inspection apparatus characterized in that the pattern inspection is performed in consideration of the following. 10. The pattern inspection apparatus according to claim 8, wherein the inspection unit configures an area based on an edge of the inspection target pattern image that has not been able to be associated, and determines the area. A pattern inspection apparatus characterized by recognizing a defect area. 11. The pattern inspection apparatus according to claim 8, wherein the inspection unit forms an area based on an edge of the inspection target pattern image that has been able to perform the association. A pattern inspection apparatus characterized by detecting an area having a non-uniform luminance distribution among the areas and recognizing the area as a defective area. 12. The pattern inspection apparatus according to claim 10, wherein the inspection unit determines a defect type based on a geometric feature amount of the defect area. 13. The pattern inspection apparatus according to claim 10, wherein said inspection means determines a defect type based on a characteristic amount relating to luminance of said defect area. . 14. The pattern inspection apparatus according to claim 8, wherein the inspection means calculates a pattern deformation amount of the inspection target pattern with respect to the reference pattern. . 15. The pattern inspection apparatus according to claim 14, wherein the pattern deformation amount includes at least one of a position shift amount, a magnification change amount, and a line width increase amount. apparatus. 16. The pattern inspection apparatus according to claim 14, wherein the inspection means adds a pattern attribute to the reference pattern. 17. The pattern inspection apparatus according to claim 1, wherein the inspection unit takes a profile on the inspection target pattern image, detects a predetermined point for each profile, and detects the detected point. A pattern inspection apparatus wherein curve approximation is performed on a point to be an edge of the inspection target pattern image. 18. A pattern inspection method for inspecting a pattern to be inspected by comparing it with a reference pattern, comprising: an inputting step of inputting an image of the pattern to be inspected; and storing an edge of the image of the input pattern to be inspected. A pattern inspection method, comprising: an inspection step of inspecting the inspection target pattern by comparing an edge of the reference pattern stored in a unit; and an output step of outputting a result of the inspection. 19. A computer-readable recording medium having recorded thereon a program for causing a computer to execute a pattern inspection method of comparing an inspection target pattern with a reference pattern and inspecting the inspection target pattern, the pattern inspection method comprising: An input step of inputting an image; an inspection step of inspecting the inspection target pattern by comparing an edge of the input image of the inspection target pattern with an edge of the reference pattern stored in a storage unit; An output step of outputting a result of the inspection.